the potential of web-based mapping and virtual … · the potential of web-based mapping and...

PII: S0198-9715(98)00014-3

THE POTENTIAL OF WEB-BASED MAPPING AND

VIRTUAL REALITY TECHNOLOGIES FOR MODELLING

URBAN ENVIRONMENTS

Simon Doyle1, Martin Dodge2 and Andy Smith3

Centre for Advanced Spatial Analysis, University College London,1±19 Torrington Place, London WC1E 6BT, U.K.

ABSTRACT. The paper will consider the potential of interactive mapping and

virtual reality technologies being developed on the World Wide Web (WWW)for the visualisation, modelling and analysis of urban environments. A range ofinnovative technologies are being developed that offer different ways of modelling

and representing built-form and associated urban information with real-timeinteraction over the Internet. We shall present a review of the capabilities of thetechnologies and how they can be applied in the field of planning and design of

urban environments. The WWW offers both an interface to, and a deliverychannel for, the built environment information as well as being a medium forlinking distributed users. We are interested in affordable ``off-the-shelf'' software

that is relatively easy to set-up and use and which requires standard PC-computing power preferable to a home user with a modem link (i.e., not high-endgraphics workstations). The advantages and disadvantages which thesetechnologies offer will be considered in terms of the level of realism and

interactivity available to the end user. Working examples of these technologieswhich are being developed by the authors will be demonstrated and discussedthroughout so as to qualify this review. The paper will also consider the

applications of these technologies in a range of contexts, such as in localplanning, urban design, development control, community participation, educationand training. The implications for a wide range of potential users including

planners, infrastructure managers, built environment students, community groupsand interested members of the general public will also be discussed. # 1998Elsevier Science Ltd. All rights reserved

137

1Corresponding author. Tel.: +44-171-391-1781; fax: +44-171-813-2843;e-mail: [email protected]; http://www.casa.ucl.ac.uk/2E-mail: [email protected]: [email protected]

Comput., Environ. and Urban Systems, Vol. 22, No. 2, pp. 137±155, 1998# 1998 Elsevier Science Ltd. All rights reserved

Printed in Great Britain0198-9715/98 $19.00 + 0.00

Pergamon

INTRODUCTION

Planners and urban designers have employed a wide variety of techniques in order toplan, model and simulate the outcome of a given process or development (Ranziger &Gleixner, 1997). Within these fields, physical models or large-scale plans have been createdwith three aims: to aid the decision making process, to democratise the planning processand to assist in the conveyance or dissemination of ideas. In the past, schemes which havesought to change given elements of the built environment have employed physical models,perhaps made of cardboard or wood based on architect's plans, perspective drawings andphotomontages in order to promote a design solution or proposed planning development.Whilst still employed, such methods may be used less extensively in the future due to theemergence of increasingly more powerful desktop computing with constantly improvinggraphics capabilities, available at modest financial cost to individuals and agencies alike.This increase in computing power is facilitating the advance of geographic informationsystems (GIS), multimedia and virtual reality (VR) software within a variety of disciplineswhich are concerned with modelling or analysis aspects of urban environments. Within aplanning context Shiffer's work is one such area where GIS has been successfully coupledto multimedia to model or depict proposed change in built environments (Shiffer, 1992,1996). The work undertaken involved the use of digital media in public participation,whereby the results of analyses within GIS were shown to an audience of decision makersor interested parties whom the analyses would affect. In a broad sense this use of GIS andrelated technologies through public participation can be seen as a precursor to the role theWorld Wide Web (WWW) could take as a distribution mechanism, not only withinplanning but in any process where consultation may be required.In the following sections, the potential of Web-oriented GIS and VR software will be

reviewed with particular reference to the disciplines of planning and urban design. Havingoutlined these potentials, the ability of the Web to serve as a dissemination mechanism willbe discussed. The discussion will be qualified by examples currently being piloted by theauthors.

EMBRACING THE WORLD WIDE WEB

In the U.K. planning community, the uptake of digital technologies is widespread, withnumerous local authorities, central government departments, utilities and policy anddecision makers utilising digital data and GIS in both day to day and strategic planning(Batty, 1998). This trend can also be seen within other countries, for instance within theEuropean Union (Von Rimscha, 1997), Australasia (Bishop, 1998) and the U.S.A. (Shiffer,1992, 1996). In this context we are interested in the way digital map data may bemanipulated and how it may be coupled to a variety of VR techniques on the WWW inorder to provide what Counsell and Phillips (1997) identify as ``a credible context forvisualising proposed changes'' which may occur within the built environment. Bysimulating change within a computed environment, information can be projected using theWWW and browser technologies as modelling and presentation platforms.Within the U.K. the exposure of GIS and digital data within decision-making agencies,

local authorities and national government agencies may be regarded as widespread yet

138 S. Doyle et al.

piecemeal in that there are a variety of systems in use across a number of agencies and local

authorities. This is perhaps an artefact of the structure or organisation of the agencies

which are responsible for the management and development of the physical environment.

Whilst the level of uptake of GIS is varied, the requirement for analysis and planning

within the urban environment is more evenly distributed. Nonetheless GIS are being used

routinely to solve problems which relate to the built environment (Longley & Clarke,

1995). From this foundation the WWW can be employed as a distribution mechanism for a

variety of GIS- or VR-based models and applications. In support of this analytical and

planning function, the WWW is beginning to be used as a support tool, examples of these

applications include Wandsworth council's on-line planning register, Devon County

Council's 2001 structure plan and Westminster City Council's Web-based planning search

engine. These examples vary in purpose and in functionality but do mirror Shiffer's (1992,

1996) work by opening up the decision-making process. It may be several years before we

see the common use of the WWW by planning departments per se but the potential is

beginning to be realised. Public expectation may also rise accordingly as the number of

connections to the WWW and the Internet increases. On-line applications currently

include banking, home shopping, publishing and distribution of information, it may be

seen as only a matter of time before planning or urban design issues are (or are expected to

be) addressed in a similar public forum (Batty, 1997a). This may be seen as an evolution of

previous desktop-based work, in that the underlying principles of collaboration and

consultation are facilitated by geographic information or GIS but with an altered delivery

mechanism (i.e., the WWW).

The following sections will review, in turn, the elements of digital technology which are

currently being harnessed and developed for Web-based dissemination. Whilst the

technologies described here may not at present be in the grasp of the majority of planners

or urban designers or to those studying the built environment due to technical or

institutional reasons, they do at present have great potential (Batty, 1997b). As part of the

authors' research these potentials have been examined in four major areas of Web-based

development;

(1) interactive mapping;

(2) solid and geometric modelling;

(3) virtual reality modelling language (VRML);

(4) multi-user worlds.

Each of these will be addressed in turn with examples of existing and potential uses being

discussed. The review progresses through several areas in which user interactivity is an

increasingly key factor leading to the point where multiple users can interact and

collaborate in computed digital environments. The paper commences with the traditional

two-dimensional representations of the urban environment and develops through the

linkage of two- to three-dimensional models and onto collaborative three-dimensional

worlds. It is envisaged that this will be regarded as a logical progression in terms of the

technologies under review. It will also serve to illustrate the development paths of the more

complex applications and will serve to illustrate how GIS and two-dimensional digital data

may be engineered in pursuit of Web-based three-dimensional modelling.

Mapping and Virtual Reality Technologies for Urban Modelling 139

INTERACTIVE MAPPING

Interactive mapping or Internet GIS has developed rapidly over the past few yearsresulting in the migration of some GIS functionality to the WWW (Plewe, 1997). Thismigration has taken elements of GIS to mass audiences. Currently there are a variety ofWeb sites which use Internet GIS for a diverse range of applications from city guides(http://www.thisislondon.co.uk/), on-line business directories (http://www.scoot.co.uk/) toenvironmental lobbying (http://www.foe.co.uk/cri/html/postcode.html). These applica-tions incorporate GIS as on-line tools to disseminate information. Whilst not planningbased, these examples show how spatial information can be used to inform a given interestgroup or audience through Internet GIS.Within the planning community there has been a steady growth in the use of digital data

of a spatial nature (Longley & Clarke, 1995). This growth is no doubt tied to the rise insophistication of GIS and the continuing migration of high-end workstation functionalityto the desktop. This migration of power to the end user and the increased availability ofdetailed digital map data has allowed professional decision makers the potential to think``spatially'' more often in analysing and visualising solutions to problems relating to thebuilt environment. Within the context of the WWW, map style images have been displayedin various formats for a number of years. Whilst these images are useful in certaininstances, they offer none of the functionality available to the experienced GIS user. Thissubsequently reduces the ability of the end user to interrogate information about a givenarea.By placing planning applications ``on-line'' (i.e., within a Web site which incorporates an

interactive map as an interface to the information), interested parties could obtaininformation on previous developments, site or land ownership, contact details in order tovoice an opinion and so on. This utilises the same technologies as the on-line tour guides orbusiness locators mentioned earlier. To date, two broad types of interactive map have beendeveloped by the commercial GIS vendors (i.e., the static map and the dynamic map;Plewe, 1997; Toon, 1997). These broad categories both employ the same kind ofclient±server relationship where the server is linked to a GIS and associated databasewhich is accessed across the Internet by an end-user's or client WWW browser. Both ofthese aspects of Internet mapping can facilitate the distribution of spatial informationrelating to the urban environment; the following sections describes dynamic mapping andstatic mapping in turn.

Dynamic Map

The dynamic map is a GIS embedded within a client Web browser wherein the user willtypically be able to zoom, pan, identify objects, hotlink to other data and turn layers ofdata on and off. User actions will be responded to in real time by a map server, thus givingthe client or end user a basic GIS interface. Within the commercial market several GISvendors have dynamic Internet mapping systems within their product range. Autodeskhave Map Guide, ESRI produce MapCafe and Intergraph and MapInfo have their ownproprietary applications. Whilst not designed to produce extensive analytical capabilitiesthese applications are empowering the Web community by making spatial data morereadily availableÐ the functionality is at a level where pre-determined data may be queriedand examined in such a way that the typical end user is not overwhelmed by unfamiliar

140 S. Doyle et al.

technology but is still capable of learning about, or contributing to any given discussionover the Web. An example of this form of Internet mapping has been developed by theauthors for strategic planning for London's Riverside NHS Trust. Within the browser, theend user is able to zoom, pan, examine attribute data, switch vector and raster layers onand off and hotlink to related Web sites and to multimedia clips. In this instance the JAVAprogramming language is being used to serve the map information to the end user. This isadvantageous in that any JAVA-enabled browser can be used to view the application(Netscape 3.0 or Internet Explorer 3.0 or later). As JAVA is platform independent the onlyprerequisite is for a suitable browser to be used at the client end. Alternatives to the JAVAapproach include the use of plug-ins, or applications which are incorporated or loaded intobrowsers at the client end, Intergraph GeoMedia and Autodesk MapGuide are examplesof the plug-in approach. The immediate disadvantage of using a plug-in is that softwarehas to be installed into the browser at the client end prior to data access. This requiressome expertise and the appropriate access permissions to alter the client configuration.Furthermore many plug-ins are platform or operating system dependent which in turn maybe restrictive in terms of accessing appropriate data.

Static Map

This kind of interactive map relies on user input to select a set of variables, which in turngenerates a map image at the server side. This map is then returned as a ``result'' of the userinput. In general, the functionality and scope of the graphic user interface are reduced incomparison to that of the dynamic interactive map but the query capabilities are usuallymore advanced. This kind of application can prove to be more complex for a ``public'' userto come to terms with if developed at too high a level. With competent programming manyGIS functions can be created according to the relevant application. On the WWW thereare a range of these query-based applications ranging from a toxin release inventory togeodemographic thematic map applications (http://www.esri.com/). It should be notedthat to date the most developed of these applications are U.S. based, this perhaps reflectsthe physical location of the major GIS vendors, it also reflects the level of accessibilitywhich exists in the U.S.A. with reference to digital data and freedom of information.Interactive mapping within the browser is a rapid area of development for the GIS vendorand user alike. With the advent of Microsoft's Internet Explorer 4.0 the desktop isprobably destined to be transformed into a quasi-browser whereby more and moreinformation and applications will be Web enabled or optimised for browser viewing.The ability to provide two-dimensional cartographic or thematic spatial information to

the Web community is indeed a useful tool and few would dispute the potential of such anapproach in distributing GIS and geographic information. Within the Web browserthough, we are able to go beyond the two-dimensional view of the world which many GISand cartographic agencies still purport. Through the use of a ``z'' dimension (height) we areable to generate models and three-dimensional models of certain urban environments. Thisallows the opportunity to provide a greater visual impact perhaps by thematically codingthree-dimensional building polygons at storey level according to any given associatedattribute (e.g., rateable value, occupancy level, housing condition, council tax valuation,building use and so on). Whilst this can be performed in a two-dimensional desktop GISthere are distinct advantages of having a three-dimensional representation especially interms of visualisation and interaction. There are also numerous ways in which otherattribute information can be manipulated by using object height as a descriptive variable


(e.g., building outlines may be extruded according to a socioeconomic or demographicattribute with colour describing some other qualitative or quantitative characteristic of theurban environment). Current research by the authors in this area is being undertakenwherein factors such as retail turnover have been generated as three-dimensional surfacesand draped over building polygons to display characteristics of English and Welsh towncentres for central government (DETR, 1998). As the majority of these three-dimensionalrepresentations can be generated as VRML files they are viewable through the Webbrowser and as such are deliverable over the Internet subject to the constraints discussedlater. This leads us to the second area of relevance in modelling the urban environment.

SOLID AND GEOMETRIC MODELLING

Solid, geometric modelling has traditionally been seated in the domain of specialistcomputer-aided design (CAD) packages running on high-end graphics workstations. Thesepackages, whilst often achieving a high degree of realism in modelling urban environments,tend to be limited to operation on single machines running the expensive proprietarysoftware, hence restricting access where available, to fixed problems or planning anddesign issues. The WWW opens up the ability to widely distribute these three-dimensionalgeometric models albeit with several caveats. The first major restriction is that models needto be developed in accordance with the limitations imposed by a low bandwidthenvironment. If dissemination is to be effective, the home user with a modem and linerental charges, should not be overlookedÐ this effectively means that download timesneed to be efficient. To achieve a wide user base on the WWW, models need to bedeveloped in a file size of not more than 300 kilobytes (K) and preferably less than 100 K.Such a limitation on file size sets the challenge of achieving high levels of realism in a lowbandwidth environment. The following sections explore issues which need to be addressedwhen modelling the built environment on the WWW. The use of VRML and otherproprietary languages will be explored in conjunction with photospatial panoramicmodelling and augmented reality (AR). The potential of each technique for modelling thebuilt environment will be explored with current examples available on the WWW.

Virtual Reality Modelling Language

VRML provides the basis for the majority of existing urban models on the WWW.Indeed, VRML is summarised in the introduction to the VRML 1.0 specification as alanguage for describing interactive simulationsÐvirtual worlds networked via the globalInternet and hyperlinked with the WWW (Bell, Parsis & Pesce, 1995). The language wasdesigned as an open standard for integrating three-dimensional worlds into the Internetand was conceived as such in 1993 through the integration of WWW extensions to SiliconGraphics Open Inventor file format, (Martin & Higgs, 1997). VRML, currently based onspecification 2.0 (Moving Worlds), has developed into a de facto standard for thedistribution of three-dimensional models on the WWW.VRML provides a flexible, cross-platform (VRML is viewable on any operating system

with the appropriate browser) environment to model urban form, the user is able to freelyexplore a model and view details from any angle, providing a very flexible way ofinterpreting a model using a suitable browser. As VRML is designed for distribution onthe WWW, a block model of a large urban environment can potentially occupy less than

142 S. Doyle et al.

100 K of file space. With an increase in complexity, scale and realism of the modelledobjects however, there comes an increase in file size. Furthermore, to achieve realismin terms of being able to identify a polygon as, an ``office block'' or an ``Edwardianterrace'', VRML modellers are sometimes forced to apply textures, or photorealisticfacades to the VRML objects in the model thus increasing file size and download time.An example of this kind of model is Virtual Tokyo, developed by Planet 9 Studios(http://www.planet9.com). Within this example the level of realism has been compromisedto achieve fluid movement on a home-based PC (taken as a Pentium 90 processor) with anacceptable download time (i.e., 90 sec using a standard 28.8 baud modem). Within thevirtual scene the end-user is presented with a series of VRML polygons which have hadtextures added to them as an aid to visual realism.The current release of VRML (VRML 2.0) allows the developer to increase the level of

achievable realism by permitting behaviours to be assigned to objects. Behaviours allow ascene to include object movement and directional sound, significantly increasing levels ofphysical realism and interactivity. The basic block model is created and appropriatetextures applied to aid user recognition. The user can be presented with the ability to moveany object within the model on the x, y or z axes. The ability to move aspects of the builtenvironment independently of one another allows each user to create their owninterpretation of any given design scenario. The use of VRML 2.0 in this manner is animportant step towards the development of Virtual Internet Design Arenas (ViDAs),which will be discussed later. However, as with VRML 1.0, VRML 2.0 is texturedependent and features such as directional sound increase the overall file size and thereforethe download time from the WWW.

TOWARDS A THREE-DIMENSIONAL GIS

There has been considerable research to push the capabilities of urban GIS from simpletwo-dimensional mapping towards three-dimensional visualisation of the built environ-ment (Faust, 1995). This has often been achieved through the linkage of CAD technologiesto GIS databases (Liggett, Friedman & Jepson, 1995). More recently there has beeninterest in using VR techniques to produce three-dimensional solid geometry models thatthe user can interactively explore and interrogate. The practical implementation of this hasbeen achieved using VRML with the three-dimensional models being viewed in separatebrowser applications. In this sense the three-dimensional GIS is created by the ``loosecoupling'' of VRML with two-dimensional spatial databases. Examples include theresearch by Martin and Higgs who coupled ARC/INFO output to VRML to visualiseurban property information (Martin & Higgs, 1997), Smith who has created a extension toMapInfo GIS called Pavan which creates VRML models (Smith, 1997) and Dodge whohas an developed scripts within ArcView GIS to produce VRML models of small-scaleurban scenes (Dodge & Jiang, 1997; Dodge, Smith & Doyle, 1997a, 1997b). Figure 1 showsthis approach of ``loose coupling'' GIS to VRML.

Adding a Third Dimension

To create a three-dimensional model from what may be called two-dimensional data,four distinct processes are applied to the data (Figure 2). In Stage 1, the conventional two-dimensional GIS database is used to map the desired spatial information. The user then


runs a program or script, usually with the GIS, that generates a VRML format outputfrom the two-dimensional geometry and attributes (Stage 2). This output is stored in a datafile, external to the GIS (Stage 3), which is loaded into a suitable VRML ``browser'' forviewing (Stage 4). The ``browser'' provides the functionality for the user to interactivelyexplore the three-dimensional model by ``walking'' and ``flying''. The user is unable tomodify the model in the ``browser''. Modification of the model requires going backto Stage 1 and making changes to the two-dimensional GIS database and then generating anew VRML model. So there is only a one-way link between the GIS and the VRML, withno means of feedback. Despite this limitation, there are several key benefits to usingVRML to produce three-dimensional models from GIS. Firstly, it is a relatively low-costoption in terms of time and money particularly as VRML browsers are largely freesoftware. As the VRML language is an agreed Internet standard and is not proprietary it istherefore easy to distribute models to the Web community.Many of the commercial GIS vendors are now realising the importance and the

possibilities of three-dimensional GIS using VR technologies. For example, ESRI Inc.have launched a powerful extension to their desktop GIS ArcView called 3D Analyst(ESRI, 1997). This will provide three-dimensional visualisation capabilities within the GISitself allowing a degree of modelling and analysis in three dimensions and the ability toexploit a VRML simile on the desktop. This release and others like it is a commercialrealisation of the potential of three-dimensional modelling and data display mechanismundertaken in the type of research undertaken by Martin and Higgs (1997) particularly inthe use of VRML. We have mentioned VR techniques and the ability to employ them bycoupling them to GISÐ the following section describes VRML and solid and geometricmodelling in greater detail.

Proprietary Web-Based Virtual Reality Modelling Software

A number of commercial vendors have developed their own proprietary modellingsoftware. One of the most widely used of these packages has been produced by Superscape(http://www.superscape.com). Superscape's VRT software has been employed extensivelyin the development of the company's Virtual World Wide Web (http://www.vwww.com)

FIGURE 1. The ``loose coupling'' of GIS to VRML within a Web browserÐ to the left is a dynamicInternet map of Whitehall and linked to it on the right is the corresponding three-dimensional VRML

model (source: authors).

144 S. Doyle et al.

and in the construction of various models of superstore developments and of tourist sitessuch as Piccadilly circus. These models are presented in a standard Web browser but withneed for a proprietary plug-in. The use of proprietary software often allows thedevelopment of innovative, highly realistic urban models on the WWW. However,the production and browsing of these models is often dependent on the purchasing ofproprietary authoring software (e.g., Superscape VRT or Silicon Graphics Cosmo suiteof applications)Ðas opposed to VRML which is an open standard. An obvious contrastbetween the open standard of VRML and the commercial alternative exists in this context.

Photospatial Panoramic Views

In contrast to the solid or block modelling discussed already, a number of softwarecompanies, have developed photospatial VR systems for use on a variety of operatingsystems. Apple's QuickTime Virtual Reality (QTVR) is characteristic of a range of WWWdevelopment tools utilising photorealistic scenes to create navigable movies. QTVR has the

FIGURE 2. The four stages of creating three-dimensional models from GIS-based data.


largest user base with over 3,000 sites on the WWW. Realism is achieved by using

photographic rendering which permits the user to pan around the scene or node. Figure 3

illustrates ``The Virtual Streetscape'' an example of the use of QTVR to model the built

environment. Developed as a demonstration project of the MIT Department of Urban

Studies & Planning's Computer Resource Laboratory, the Virtual Streetscape illustrates

how 360 panoramic views can be linked to aerial photographs to create a linkage between

more traditional, two-dimensional images and three-dimensional VR scenes.

Whilst photographically rendered scenes are highly realistic, the user's viewpoint is fixed

to the original location of the camera. The user is not able to ``walk'' around the scenes in

the same way as in VRML though nodes can be linked allowing users to ``jump'' between

scenes creating a sense of movement. Typical file sizes for QTVR are 200±300 K for each

node, placing a restriction on movement as each new node is downloaded separately as the

user moves around. The development of a WWW interface featuring panoramic scenes

may be viewed in four distinct stages.

Stage 1 consists of the data collection for each of the scenes to be created. Each scene

consists of approximately 14 photographs, depending on the camera lens used. The

photographs are taken, in this case using a digital camera, allowing each to overlap by

approximately 5%. The images are then ``stitched'' together (Stage 2) to produce a full 360

panorama using specialist software such as Photovista (http://www.livepicture.com). Links

to other scenes and Web resources can now be added (Stage 3). Once the panoramas have

been produced they can then be integrated into a WWW page (Stage 4). The ``Wired

Whitehall'' integrates information relating to the built environment through photospatial

scenes using hot linking software by Jutvision (http://www.visdyn.com). The use of

photorealistic VR scenes on the Web is increasing dailyÐ in order to add extra value to

FIGURE 3. ``Virtual Streetscape'' (Computer Resource Laboratory, Department of Urban Studies &Planning, MIT).

146 S. Doyle et al.

the scenes within a planning context, representations of real world objects can be placedwithin the panoramas. This object placement augments the actual with the proposed thusallowing an end-user to simulate or visualise proposed change in a specific location. Thistechnique is encompassed within the broad term of AR (Tuceryan et al., 1995). AR is atechnology in which a user's view of the real world is enhanced or augmented withadditional information generated from a computer model. The enhancement may take theform of labels, three-dimensional rendered models or shading modifications. The uses ofAR in modelling the built environment are numerous, ranging from the placement of apiece of street furniture into a photorealistic scene to the overlay of new urban form. Toa limited extent, ``Wired Whitehall'' is a basic example of AR, in that virtual billboards arecreated within photorealistic panoramas, in order to create hyperlinks of relevance to theurban scene. For example, a virtual billboard near Downing Street allows access tothe ``No. 10 Downing Street'' Web site. A degree of interactivity within the urban scene isdesirable if various planning schemes are proposed. This can be achieved by combiningVRML 2.0 objects and photospatial scenesÐAR on the WWW.Figure 4 illustrates an application developed by the authors where a VRML 2.0 object is

placed into a 360 panoramic image using RealVR. RealVR (http://www.livepicture.com)is a plug-in for the Netscape browser. Whilst similar to QTVR, RealVR allows theinsertion of VRML 2.0 objects into scenes. Once an object has been placed, augmentingreality, the user is able to move the object into and out of the panorama. AR is an importantstep forward in the ability to interact and model the built environment on the WWW.

VIRTUAL WORLDS

Having discussed ways in which the built environment can be modelled on the WWW, itis useful to explore the concept of virtual worlds and their ability to introduce aspects ofinteraction and social behaviour into the types of models already discussed. By allowing

FIGURE 4. Augmented reality on the WWW: object placement within a panoramic scene.


movement within the models we create, users can be presented with the chance to becomeinvolved with planing problemsÐ for planners this facilitates consultation and also allowsthe planning community to interact in a common digital space in the pursuit of a design orplanning solution.Within virtual worlds participants are generally represented in the model by an avatar or

digital alter ego. A user's avatar may take any form, although they are usually based onhuman form. The user is able to control their avatar, which will dynamically update theirview of the world and the other participants' avatars. The user is able to see the three-dimensional world through the eyes of their avatar. This allows the user the ability to turnaround and look at a fellow avatar, whist holding a conversation thus introducing ameaningful level of social interaction. The more advanced avatar systems available on theWWW also allow the use of physical personal gestures. The most advanced virtual worldson the WWW have the following features (Rockwell, 1997):

(1) insert/delete objects (e.g., avatars) in scenes at run-time;(2) merge multiple sound streams from distributed sources into the shared scene's

current ambient sound (e.g., voices over music);(3) track and communicate the state/behaviour of objects in real time;(4) allow (sets of) objects to be ``driven'' by users in real time;(5) let imported objects become persistent;(6) protect the scene from damage by imported objects;(7) assign objects to a series of different ``owners'';(8) support persistent roles (for people) and rules (for scenes);(9) link objects dynamically to external data/functions;(10) support the free exchange of information among objects.

Within the realm of planning and decision making in the urban environment (or eventhe rural environment) these virtual worlds offer the opportunity to practice, simulate andvisualise a design or planning issue in real time in a ``dry'' environment. All that is basicallyrequired is a model (a replication of a ``real-world'' space) and the ability to serve it tointerested clients. A schema of how a virtual worldmay be organised is described in Figure 5.

Virtual WorldsÐThe Five Stages of Development to Full Interactivity

The creation of virtual worlds on the WWW and their ability to be used forcollaborative modelling and design of the built environment may be viewed in five distinctstages, ranging from the basic HTML (Hyper Text Mark-up Language) Web site to a fullydeveloped three-dimensional collaborative simulated urban environment. Table 1 andFigure 6, adapted from Rockwell (1997), illustrate the five stages to full interactivity invirtual worlds.Current technology lies between Stages 3 and 4, with the development of open-based

standards for the development of multi-user virtual environments. There are more than 10different software vendors developing multi-user virtual worlds (Yahoo, VR listing). Thereis considerable variation in their quality and functionalityÐ two advanced examples arethe Blaxxun and Activeworlds environments.The Blaxxun virtual environment client, CCPRO, is a free Web browser plug-in which

can be downloaded from the Blaxxun Web site (http://www.blaxxun.com/); however, it isonly available for Windows 95/NT platform. The client comprises three elements, the

148 S. Doyle et al.

FIG

URE5.Virtualworldsclient/server

model.


three-dimensional view, a chat window and a control panel. The three-dimensional aspect

of the system utilises VRML versions 1.0 and 2.0. The ability to interact with objects

within a multi-user three-dimensional space allows the creation of what we have termed

``Virtual Internet Design Arenas'' (ViDAs). The Blaxxun CCPRO client also supports

environments created using Superscape's proprietary modelling language. The second

example is the ActiveWorlds browser (http://www.activeworlds.com). This is a stand-alone

client application, again only available for Windows 95/NT platform. The client has a

large three-dimensional view window, plus a control panel and the ``chat'' window. The

largest of the ActiveWorlds virtual worlds is called Alphaworld which currently has over

200,000 registered ``residents'' who are able to build objects and structures using a simple

``cut-and-paste'' system placement process. Such activities allow the production of ``land

use'' maps of Alphaworld, illustrating virtual urban sprawl in the digital environment.

Over a period of 15 months over 10 million building objects were placed on the digital

FIGURE 6. Development time line of the five stages towards full interactivity for a virtual world.

Table 1. The Five Stages Towards Full Interactivity

Stage of evolution Description Level of use for collaborativeeducation/design

HTML Web site No community development WeakÐonly information availableHTML site with ``chat'' ability (i.e.,Internet relay chat or Java-basedchat)

Simple anonymous text chat MediumÐ interaction, ability tocommunicate in real-life

Avatar-based graphical interface Personal, organised and structuredinteractions

StrongÐdevelopment of design/education arenas based on propertysoftware

Avatar-based 3D using openstandards and interactions

Seamless interconnection ofcommunities and sharing ofknowledge across communities

Very strongÐdevelopment ofworldwide design/education arenas

Avatar-based 3D/open standards/real-time audio synchronisation

Seamless interconnection ofcommunities with naturalcommunication

Very strongÐdesign/educationarenas further enhanced with real-time audio

150 S. Doyle et al.

plain (Damer, 1998). The map provides ``teleport'' access to any area of Alphaworldsimply by clicking on the desired location.

THE ``VIRTUAL DESIGN ARENAS'' INITIATIVE

At CASA we are currently developing our own virtual world server, on which we canhost small three-dimensional design scenarios. This will be used for experimental teachingand modelling of urban design and planning in a ViDA. The ViDA initiative will be acollaborative venture between the VENUE and On-line Planning projects, both ongoing atCASA. These projects deal with the linkage of GIS to urban design processes andvisualisation (VENUE being Virtual Environments for Urban Environments) and the roleof the WWW as a medium for planning support or dissemination (On-line Planning).Within the proposed ViDA initiative participants may include:

(1) urban design and planning students;(2) interested academics;(3) professional planners and urban designers;(4) ordinary people in public participation experiments;(5) planners and urban designers;(6) the Internet ``general public''.

By developing specific virtual worlds, design scenarios can be run in which variouspeople can participate. Two distinct categories of design scenarios ``Virtual Real Places''and ``Real Virtual Spaces'' will be facilitated. The first type, ``Virtual Real Places'' refers toscenarios using three-dimensional models of real world places (e.g., real developments sitesin London). Whilst ``Real Virtual Spaces'' will be completely fictional models, not relatedto any particular real place. We believe that the three-dimensional worlds do notnecessarily have to be based on physically real places to be useful in the design andplanning process. In many respects it may be easier to represent the ideas of good and baddesign in truly virtual spaces.Figure 7 illustrates the basic design process of the virtual world. Initially ViDA users will

be asked to explore the potential of the computer-mediated virtual world, (e.g., becomingused to moving through the three-dimensional model world or using the ``chat'' tool tocommunicate). Users will then be presented with a series of virtual worlds to evaluateaccording to a set criteria. Eventually they will be presented with virtual worlds where theycan modify the three-dimensional models. In this instance a range of powers of``modification'' can be permitted such as changing building colours or textures, alteringsize and plot position, or removing buildings and objects (like street furniture), to creatingcompletely new structures and layouts. Permissions can also be set in order to determinewho has the power to make certain changes to the model (development control). All useractivities in the multi-user worlds (avatar movement, action and ``chat'' communications)will be discretely logged on the server, thus creating extensive metadata. Analysis of thisactivity log plus conventional oral and written feedback, will provide interesting insightsinto group evaluations, design and working in virtual environments. The virtual lecturetheatre will be used as a base for the users to initially meet in the virtual world and beintroduced to the design concepts by a virtual lecturer.


It should be stressed that the ViDA initiative is an experiment in using available multi-

user world technologies in urban design. However, we have been considering some of the

possibilities which may open up if the initial experiment proves the notion that virtual

design and planning are viable (e.g., the possibility to construct more elaborate scenarios,

with ``role play'' where different participants assume the ``roles'' of the planner, the

developer, the local resident). We would also like to explore the possibilities of

``embedding'' rules into the virtual world of what can be built where. In this way,

electronic planning constraints can be implemented and enforced on the designers. In some

senses we would be adding a measure of intelligence to our three-dimensional world. We

would also envisage the linkage of attribute data to buildings and objects in our worlds

perhaps via an object-oriented database where the object identifier forms a primary key.

The communication between participants could also be improved beyond ``chat'' text

messages using desktop video-conferencing and ``whiteboard'' technologies.

There are obviously limitations in the application of virtual worlds to urban design and

planning issues. The realism required in the three-dimensional models to be effective for

our purposes will needs to be fit for purpose and there may be scenarios where building

textures are required or alternative scenarios where building roof morphology need to be

calculated accurately. Another limitation in the three-dimensional modelling formats used

by the multi-user world systems already outlined is the lack of shadows. Clearly, this will

hamper the realism we can achieve in modelling the built environment where lighting is an

important issue in a design. Although the clients for the multi-user world system are

generally free, they are mostly exclusive to the Windows 95/NT platform. This is a

limitation as many users in universities are using systems with older versions of Windows

or Macintoshes. Furthermore, server software has to be bought, costing from a few

FIGURE 7. The basic virtual design process of ViDA.

152 S. Doyle et al.

hundred pounds to a couple of thousand pounds sterling. There is also the danger ofbecoming locked into proprietary software systems, over which authors have littledevelopment control, which can leave users at the mercy of the vendor. This can beparticularly problematic with regard to the smaller vendors of Internet software whichtend to have unpredictable life-spans. The communication between participants(particularly ones at geographically separate locations) is also dependent on typing textmessages in a ``chat'' window. This can be problematic for carrying on in-depth, lengthy``conversations'', especially for people not adept at typing. However there are manybenefits to pursuing the virtual world as a teaching or visualisation medium, not least theconnectivity offered by the WWW, in which these applications are founded. The interfacesto these user worlds are also relatively straightforward to master in terms of navigationand communication, with cursor keys or the mouse providing the control mechanism.Model creation is also becoming less complex following research in the linkage of two-dimensional data in GIS to three-dimensional VRML or CAD output alluded to inprevious sections (Dodge & Jiang, 1997). The models can be modified over a series ofhours and readily placed on-line.

CONCLUSIONS

This paper has reviewed Web-based technologies available and suitable for the plannerand urban designer. The several fields have been discussed in terms of their applicability toWeb-based applications relating to the urban environment, these being interactivemapping, solid, geometric modelling, photospatial panoramic views, AR and multi-uservirtual worlds. The opportunity these elements allow in order to practice, simulate,visualise and conceptualise issues which relate to the urban environment have also beendiscussed but there are several issues arising from the potential on offer, especially in termsof current and future Web-based urban modelling. Firstly there are various restrictions inplace in relation to commercial datasets and their distribution on the WWW. There arealso restrictions in place in terms of confidentiality and data protection. The researchwhich has been and will continue to be undertaken by the authors acknowledges this. TheInternet mapping applications developed and described herein are all subject to theseconstraints relating to commercial copyright (e.g., Ordnance Survey data) and institutionalsensitivities (e.g., Riverside NHS Trust) thus curtailing who outside the local area networkcan view the applications. Furthermore some data types do not lend themselves readily tothe creation of three-dimensional models. Building polygons are difficult to extract fromremotely sensed images of dense urban areas in comparison to detailed survey data. Roofmorphology is also difficult to readily obtain from most commercial three-dimensionaldatasets. Notwithstanding these factors there is currently tremendous activity in theapplication of three-dimensional visualisation and linkage of GIS to VRML in the vendorand academic communities.In terms of infrastructure there is also a limitation in terms of band-width in displaying

some models live over the WWW. The major casualties of this restriction are larger (i.e.,2 MB+) VRML models. Internet Map servers, Activeworlds and photo-spatialpanoramas can all be reduced in file size by the authors to present a real-time interfaceto the relevant information. There are also costs and benefits in adopting differingsoftware to implement the various techniques of data distribution outlined already. Whilst


there is a large amount of free software or trial shareware on offer, large organisationssuch as local authorities tend to discourage the unstructured use of such applications dueto corporate IT policy or for security reasons. There are also resource issues in terms ofstorage space or hardware power and currency. Fortunately most if not all of theapplications in this paper are capable of being run in a modest NT environment and assuch provides a relatively low cost mechanism for development and dissemination. Interms of human factors or institutional issues there is still the need to convince people ofthe relevance and usefulness of embracing the digital media which are potentially powerfultools though this is common in any field where new technology is being introduced. Thisreview paper promotes the case for the integration of digital data, GIS, VR and the Web interms of opening up elements of planning processes. There is a huge potential forcombinations of the above technologies in urban decision making and with regard to theinvolvement of the wider community in the decision-making process. Whilst some issueshave been summarily reviewed, it is anticipated that the relevance of the emergenttechnologies and the WWW will have been made clearer in terms of data and modeldelivery and distribution with regard to the built environment.

REFERENCES

Batty, M. (1997a). The Computable City. International Planning Studies, 2, 155±173.Batty, M. (1997b). Virtual Geography. Futures, 29, 337±352.Batty, M. (1998). Digital planning. In P. K. Sikdar, S. L. Dhingra, & K. V. Krishna Rao (Eds.), Computers in

urban planning and urban management keynote papers (pp. 13±30). New Delhi: Narosa.Bell, G., Parsis, A., & Pesce, M. (1995). VRML: The virtual reality modelling language version 1.0 specification.

In VR for the people: A brief history of VRML [Online]. Available: http://www.gnn.com/gnn/wr/sept29/features/vrml/history.html

Bishop, I. D. (1998). Planning support: Where is the system?ÐA review of developments in hardware andsoftware. In P. K. Sikdar, S. L. Dhingra & K. V. Krishna Rao (Eds.), Computers in urban planning and urbanmanagement keynote papers (pp. 37±49). New Delhi: Narosa.

Counsell, J., & Phillips, B. (1997). Appropriate data for navigable 3D cityscape interfaces to urban informationsystems. In S. Hodgson, M. Rumor, & J. J. Harts (Eds.), Proceedings Third JEC-EGI Vienna (pp. 298±307).Netherlands: 105 Press.

Damer, B. (1998). Avatars! Exploring and building virtual worlds on the Internet. Berkeley, CA: Peachpit Press.DETR (1998). Town centres: Defining boundaries for statistical monitoring, feasibility study. London: HMSO.Dodge, M., & Jiang, B. (1997, November). Geographical information systems for urban design: Providing new

tools and digital data for urban designers. Paper presented at the Learning Spaces Conference, De MontfortUniversity, Milton Keynes [On-line]. Available: http://www.casa.ucl.ac.uk/publications/learning_spaces/

Dodge, M., Smith, A., & Doyle, S. (1997a). Urban science. GIS EUROPE, 6(10), 26±29.Dodge, M., Smith, A., & Doyle, S. (1997b). Visualising urban environments for planning and urban design.

AGOCG Graphics, Visualization and the Social Sciences Workshop (Report No. 33).ESRI (1997, Spring). Introducing ArcView 3D Analyst: Powerful tools for creating, analyzing, and visualizing

data in 3D on the desktop. ARCNews [On-line]. Available: http://www.esri.com/base/news/arcnews/spring97articles/07-3d.html

Faust, N. L. (1995). The virtual reality of GIS. Environment and Planning B: Planning and Design, 22, 257±268.Liggett, R., Friedman, S., & Jepson, W. (1995). Interactive design/decision making in a virtual urban world:

Visual simulation and GIS. In Proceedings of the 1995 ESRI User Conference. California: ESRI.Longley, P., & Clarke, G. (Eds.). (1995). GIS for business and service planning. Cambridge, UK: GeoInformaton

International.Martin, D., & Higgs, G. (1997). The visualization of socio-economic GIS data using virtual reality tools.

Transactions in GIS, 1(4), 255±265.Plewe, B. (1997). GIS online: Information, retrieval, mapping and the Internet. Santa Fe, NM: OnWord Press.Ranziger, M., & Gleixner, G. (1997). GIS-datasets for 3D urban planning. In S. Hodgson, M. Rumor, & J. J.

Harts (Eds.), Proceedings Third JEC-EGI Vienna (pp. 298±307). Netherlands: 105 Press.Rockwell, B. (1997). From chat to civilization: The evolution of online communities [On-line]. Available: http://

www.blaxxun.com/company/vision/cmnty.html

154 S. Doyle et al.

Shiffer, M. J. (1992). Towards a collaborative planning system. Environment and Planning B, 19, 709±722.Shiffer, M. J. (1996). Community-building communications technologies and decision support systems. MIT

Colloquium on Advanced Information Technology, Low-income Communities and the City. Massachusetts,USA: Massachusetts Institute of Technology, Department of Urban Studies and Planning.

Smith, S. (1997, October). A dimension of sight and sound. Mapping Awareness, pp. 18-21.Toon, M. (1997). The world by your window. GIS EUROPE, 6(11), 38±41.Tuceryan, M., Geer, D. S., Whitaket, R. T., Breen, D., Crampton, C., Rose, E., & Ahlers, K. H. (1995).

Calibration requirements and procedures for augmented reality (No. ECRC-95-14). Purdue University,Indianapolis: European Computer-Industry Research Centre.

Von Rimscha, M. (1997). 3D or not 3D? GIS EUROPE, 6(10), 20±22.

APPENDIX A: WEB RESOURCES

ActiveWorlds (http://www.activeworlds.com/)Blaxxun (http://www.blaxxun.com/)ESRI (http://www.esri.com)On-line Planning (http://www.plannet.co.uk/olp/)Hot links to the virtualworlds industry (http://www.ccon.org/hotlinks/hotlinks.html)

Virtual Streetscape (http://yerkes.mit.edu/ncpc96/home.html)


the potential of web-based mapping and virtual … · the potential of web-based mapping and...

Documents